
A genome-wide CRISPR screen reveals novel drivers of Enzalutamide resistance in prostate cancer
Abstract
Therapeutic resistance to androgen receptor inhibitors, such as enzalutamide (EZ), is a major cause of mortality for prostate cancer (PC) patients. Resistance can occur through lineage plasticity, in which PC cells transdifferentiate from an androgen receptor (AR)-dependent to an AR-independent neuroendocrine (NE) cell type. However, the mechanisms that govern therapy resistance are not well defined. Molecular heterogeneity across patients suggests multiple pathways to resistance acquisition exist. This thesis aimed to provide a resource of molecular drivers of EZ resistance and to identify novel targets to treat EZ-resistant PC. Herein, we performed a genome wide CRISPR knockout screen in LNCaP PC cells treated with EZ. This screen identified gene loss events characterized into four functional categories: 1) Essential, 2) Pro-proliferative, 3) EZ-resistant and 4) EZ-sensitive genes. To investigate screen hits in advanced disease, LNCaP cells double knockout (DKO) for RB and p53 were generated, as RB-p53 co-deletion is common in resistant patients. DKO cells are EZ resistant and displayed increased expression of stemness and NE genes, by RNA sequencing. An overlap of screen EZ-sensitivity genes and DKO up-regulated genes from RNA sequencing identified the transcription factor HOXA9. Depletion or chemical inhibition HOXA9 sensitized cells to EZ, whereas its overexpression drove EZ resistance. Our screen identified the transcriptional repressor TBX18 as an EZ-resistant gene. TBX18 mutation promoted EZ resistance and is associated with poor prognosis in PC patients. The transcription factor SOX4 scored as an essential gene in our screen. SOX4 deficiency hindered cell growth, whereas SOX4 overexpression in DKO cells drove EZ resistance. Finally, we investigated the applicability of HOXA9 as a biomarker of EZ resistance. Immunohistochemical staining for HOXA9 and its downstream target genes SOX4 and FLT3 were optimized and validated on prostate tumour specimens. Positive and negative expression in tumours sections were observed for each marker. Altogether, our CRISPR screen provides a resource of molecular mediators of EZ resistance in PC. We validated several novel drivers that can potentially serve as biomarkers and molecular treatment targets. Currently, no targeted therapeutics are available for resistant patients, so there is a need to identify novel treatment strategies to combat this disease.